Requirements

The following libraries are required:

• torch
• paths
• nn

Files

The file mnist.netcontains the neural network and has to be in the same directory as classify.lua

Execution

> require 'classify'
true
> mnist_set = torch.load('mnist.t7/test_32x32.t7', 'ascii')
> testData = mnist_set.data
> res = getLabels(testData)
[================================ 10000/10000 ===============================>]

Data format

The input tensor has to be a torch.ByteTensor of size n,1,32,32

The output tensor is a torch.ByteTensor of size n